Clinical trials for registration in the European Union: the EMEA 5-year experience in oncology.

The evaluation of quality, safety and efficacy of medicinal products by regulatory agencies is a necessary step for obtaining marketing authorisation in the European Union (EU). The objective of this paper is to outline some key aspects of the EU regulatory system relevant to the field of oncology. Regulatory standards in oncology and the experience with anti-neoplastic and endocrine therapy agents in the centralised EU review system are presented. Also, the most common pitfalls encountered with oncology applications that were rejected are illustrated. The review is based on the first 5 years (1995-1999) of operation of the European Agency for the Evaluation of Medicinal Products (EMEA). Nineteen out of 30 different oncology drugs submitted during this period were approved, based on the recommendations of the Committee for Proprietary Medicinal Products (CPMP), which is the main scientific body within the EMEA responsible for evaluating medicines for human use. Although, in general, randomised trials to test the benefit of the new drug are a prerequisite for approval, this series showed that in strictly defined situations, approval could be obtained based on non-randomised trials measuring surrogate endpoints, provided the applicant agreed to the completion of a further programme of studies. These situations were the majority in our series and the CPMP has produced a guideline in oncology to address these specific requirements. Eleven out of the 30 oncology applications did not establish a positive benefit/risk profile of the drug in the proposed therapeutic indication, and were therefore rejected. Failure generally occurred because applications were based on too small series of patients or relied on the results of exploratory analyses after pre-defined analyses had failed to produce the expected results or, particularly for diagnostic agents, due to the fact that the effect of the agent had been measured only in terms of endpoints that were of unclear relevance.